Replenit
Hire Maestro

Glossary · Output Unit · Retail AI

What is a Golden Decision Event?

By Emre Erdogan, Head of AIPublished July 16, 2026
Definition

A Golden Decision Event is Replenit's execution-ready output unit. It packages a 1:1 customer decision (what should happen, for which customer, why now, with which product, in which tone, toward which commercial outcome) as a portable, inspectable JSON structure ready to fire into any CRM, CDP, or marketing-automation platform.

Most systems output signals that still need a human to act on them: a score to interpret, an audience to build a campaign for, a dashboard to read. A Golden Decision Event is different. It is the finished decision itself, sealed as a structured object that carries everything needed to execute: the customer, the action, the product, the timing, the tone, the intended outcome, and the reasoning behind it.

Because it is a portable, inspectable JSON structure, a Golden Decision Event is execution-ready across 120+ platforms. It fires straight into the CRM, CDP, or marketing-automation tool a retailer already runs, so the decision Maestro commits is the same object that gets delivered, with nothing lost in translation.

The distinction

A committed decision, not a signal

A score is a number and an audience is a list. A Golden Decision Event is a decision with its rationale attached.

A score ranks likelihood and leaves the decision to someone else. An audience is a list of customers that still needs a campaign built around it. Both are inputs to a decision, not the decision itself, and both require human work before anything reaches a customer.

A Golden Decision Event is the committed decision. It states what should happen for a specific customer, why now, and toward which outcome, with the rationale traveling alongside so it can be inspected and trusted. And it is execution-ready, not a suggestion waiting on a marketer to assemble it.

Under the hood

Inside a Golden Decision Event

An abbreviated example. The decision, its rationale, and its delivery targets travel together as one object.

json
{
  "event": "golden_decision_event",
  "customer": {
    "id": "#31205",
    "lifecycle_stage": "post_purchase"
  },
  "decision": {
    "workflow": "cross_sell",
    "action": "recommend_complementary_product",
    "why_now": "Bought a straightener; no heat-protection or repair in history",
    "product": { "sku": "OLA-N7-30", "name": "Olaplex No.7 Bonding Oil" },
    "tone": "expert, reassuring",
    "channel": "email",
    "outcome_target": "cross_sell_revenue"
  },
  "rationale": "Frequent heat styling without protection causes visible damage; bond repair closes a protective gap in this customer's routine.",
  "confidence": 0.87,
  "ready_for": ["Klaviyo", "Insider", "Braze", "+120 platforms"]
}

Marked up as inspectable JSON so both execution platforms and language models can read the decision and its reasoning directly.

Side by side

Score, audience, or decision?

Three things a system can hand you. Only one is ready to execute on its own.

A number

Score

What it is
A probability or rank
Decision committed?
No
Execution-ready?
No: needs interpretation
A list

Audience

What it is
A list of customers
Decision committed?
No
Execution-ready?
No: needs a campaign built
A decision

Golden Decision Event

What it is
A committed 1:1 decision with rationale
Decision committed?
Yes
Execution-ready?
Yes: across 120+ platforms
16.6X
return on investment in 54 days

Proof · iBOOD

iBOOD went from zero retention to a 6.3% revenue share in 54 days by firing Golden Decision Events straight into its CRM, reaching a 16.6X return on investment.

Read the iBOOD case study

FAQ

Common questions about the Golden Decision Event

It packages a 1:1 customer decision: what should happen, for which customer, why now, with which product, in which tone, and toward which commercial outcome, plus the rationale behind it. It is expressed as a portable, inspectable JSON structure.

Keep exploring

Related terms

See Golden Decision Events in the platform